Model Reference Sliding Mode Control for RPMTM with Neural Network Load Torque Observer

Unpredictable plant parameter variations, external load disturbances and nonlinear dynamics which exist in ring permanent magnet torque motors (RPMTM) seriously deteriorate the drive performance of system at low speeds. A model reference sliding mode control scheme which features good robustness against parameter variations is proposed in this paper firstly. Then a neural network load torque observer is presented to observe and compensate external load disturbances. The analysis, design and simulation of the proposed model reference sliding mode control scheme controller and neural network load torque observer are described. Simulation results show that good control performance, both in the command-tracking and the load-regulating characteristics of the rotor position, is achieved.